Ion Crystal Transducer for Strong Coupling between
Single Ions and Single Photons
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Citation
Lamata, L. et al. “Ion Crystal Transducer for Strong Coupling
Between Single Ions and Single Photons.” Physical Review
Letters 107.3 (2011) © 2011 American Physical Society
As Published
http://dx.doi.org/10.1103/PhysRevLett.107.030501
Publisher
American Physical Society
Version
Final published version
Accessed
Mon May 23 10:56:53 EDT 2016
Citable Link
http://hdl.handle.net/1721.1/67347
Terms of Use
Article is made available in accordance with the publisher's policy
and may be subject to US copyright law. Please refer to the
publisher's site for terms of use.
Detailed Terms
PRL 107, 030501 (2011)
week ending
15 JULY 2011
PHYSICAL REVIEW LETTERS
Ion Crystal Transducer for Strong Coupling between Single Ions and Single Photons
L. Lamata,1,* D. R. Leibrandt,2 I. L. Chuang,2 J. I. Cirac,1 M. D. Lukin,3,4 V. Vuletić,2 and S. F. Yelin3,5
1
Max-Planck-Institut für Quantenoptik, Hans-Kopfermann-Strasse 1, 85748 Garching, Germany
Center for Ultracold Atoms, Department of Physics, MIT, Cambridge, Massachusetts 02139, USA
3
ITAMP, Harvard-Smithsonian Center for Astrophysics, Cambridge, Massachusetts 02138, USA
4
Department of Physics, Harvard University, Cambridge, Massachusetts 02138, USA
5
Department of Physics, University of Connecticut, Storrs, Connecticut 06269, USA
(Received 24 February 2011; published 11 July 2011)
2
A new approach for the realization of a quantum interface between single photons and single ions in an
ion crystal is proposed and analyzed. In our approach the coupling between a single photon and a single
ion is enhanced via the collective degrees of freedom of the ion crystal. Applications including singlephoton generation, a memory for a quantum repeater, and a deterministic photon-photon, photon-phonon,
or photon-ion entangler are discussed.
DOI: 10.1103/PhysRevLett.107.030501
PACS numbers: 03.67.Bg, 42.50.Dv, 42.50.Ex
Realization of efficient quantum interfaces between
single photons and single matter qubits is one of the
most important and challenging goals in quantum information science [1]. It enables a wide variety of potential
applications ranging from scalable quantum computing
schemes to quantum networks and single-photon nonlinear
optics [2]. Much progress has been achieved towards the
realization of such quantum information implementations
over the past decade [3–18]. Most of these have been based
on neutral atoms, where quantum states can be stored for
long times [3–15]. The realization of a quantum optical
interface for isolated single ions, which are among the
most promising qubit candidates [19], is still an outstanding challenge as the achievable coupling strength is typically small under realistic experimental conditions [20,21].
In this Letter, we propose a technique to collectively
enhance the coupling between single photons and single
ions using ion crystals. We consider a linear chain of ions
inside an optical cavity [see Fig. 1(a)]. Strong coupling
between a single ion and a single photon is realized by
collective enhancement of the coupling
pffiffiffiffiof the photon with
an ensemble of N ions, given as g0 N , where g0 is the
vacuum Rabi frequency, i.e., the coupling between a single
ion and the incoming photon that enters the cavity. This
means that the photon state can be mapped onto a collective internal excitation of the ions in the absorption process. Subsequently, this state will be transferred to a
phonon, i.e., a motional mode of the chain. Finally, the
phonon state will be mapped to a single-ion state. Since the
latter two couplings can in principle be accomplished
with
pffiffiffiffi
arbitrarily strong laser fields, the collective g0 N coupling
can dramatically improve the overall fidelity of singlephoton single-ion coupling.
We will show that using this mechanism it is possible to
coherently transfer with high fidelity an arbitrary internal
state of a single ion onto a single photon exiting the cavity,
or an arbitrary state of a single incoming photon onto an
0031-9007=11=107(3)=030501(4)
internal ionic state. This can be used for quantum coupling
of single-ion qubits in distant cavities or, alternatively, for
nonlinear quantum operations (quantum gates) between
single-photon qubits. In what follows we present an analysis taking into account the inhomogeneous spacing of
the ions in a linear trap. This resulting inhomogeneous
coupling between the cavity photon and the ions has a
sizable effect on the fidelities, and we suggest an approach
for the phonon-collective internal transition that will
mimic the collective internal-photon transition. The total
fidelity thus can be maximized by having the same relative
coupling constants on each ion for both transitions.
First, we briefly outline our approach for achieving
strong single-photon single-ion coupling in an optical cavity with intermediate coupling to a single ion. The system
consists of a string of N ions with a -level internal
structure as shown in Fig. 1(b), i.e., with ground states
j0i and j1i and an excited state jei. g0 is the vacuum Rabi
frequency on the j0i ! jei transition, and the transition
j1i ! jei is driven by a classical field with Rabi frequency
1 and homogeneous coupling to all ions. We also assume
a quadrupole electric transition between the j0i and j1i
states, with Rabi frequency . The collective internal ion
excitation consists then
P of the Dicke-like states j0i ¼
j01 02 . . . 0N i and j1i / i gi j01 02 . . . 1i . . . 0N i, etc., which
takes into account the inhomogeneous coupling of the
ions to the cavity field with coupling coefficients gi ¼
g0 sinðkz0i Þ. Here gi is the coupling of the cavity photon
to ion i, where z0i is the equilibrium position of the ion,
and k the photon wave vector. The single-ion excitation is a
metastable state j10 ii , where j10 i can be the same or different from j1i. In our approach, in step I we map the probe
photon onto the collective ion excitation, j0i ! j1i, via a
stimulated Raman adiabatic passage (STIRAP) process
[22]. p
The
ffiffiffiffi probe transition is thus collectively enhanced
by g0 N [12]. Step II consists of coupling this state to a
phonon j1; npn ¼ 0i ! j0; npn ¼ 1i [Fig. 1(c)], where npn
030501-1
Ó 2011 American Physical Society
PRL 107, 030501 (2011)
week ending
15 JULY 2011
PHYSICAL REVIEW LETTERS
FIG. 1 (color online). Schematic of the setup. (a) Chain of N
ions in cavity with incoming photon and two pairs of off-axis
lasers comprising a quadrupole field. (b) Ionic level scheme. States
j0i and j1i are metastable states (e.g., 3D5=2 and 4S1=2 in 40 Caþ ),
and jei is an excited state (e.g., the 4P3=2 state in the same ion). For
step I, the transition j0i ! jei is coupled
by a single photon with
pffiffiffiffi
the effective Rabi frequency g0 N . The 1 laser is directed
perpendicularly to the chain. For step II, the states j0i and j1i
are coupled by a quadrupole transition. (c) Coupling to a phonon
mode. The harmonic energy level of the phonon excitations, e.g.,
of the c.m. mode, is denoted for different internal states. The singly
excited phonon mode can be addressed by a transition frequency
that is red detuned from the bare ionic transitions by the frequency
of the phonons !pn by a quadrupole transition. The laser scheme
for this part is depicted in (a).
denotes the number of phonons in a selected mode, e.g.,
axial center of mass (c.m.), via an adiabatic passage process, by using the quadrupole transition coupling j0i and
j1i. In step III, the phonon is mapped onto a single ion j,
i.e., j0; npn ¼ 1i ! j01 . . . 1j . . . 0N ; npn ¼ 0i. The main
limitation to this schemepisffiffiffiffithe compromise between large
N for enhancing the g0 N coupling strength of the first
step, and not too large N such that the coupling to the
phonons, scaling inversely with the mass of the system, is
not reduced. We find about 40 ions to be both experimentally feasible and providing sufficient coupling on both
transitions. Thus single-photon single-ion coupling is
achieved.
The details of the scheme are most easily understood by
describing the time reversed process, so we start by analyzing how to map a single-ion state onto a phonon state
(step III).
In order to map the state of the ith ion, we propose to use
a Raman transition tuned to the red sideband, see Figs. 1(b)
and 1(c), from j01 02 . . . 1i . . . 0N ; npn ¼ 0i to j0; npn ¼ 1i.
This process can be achieved with fidelity F0 ¼
jh c f j c ðtÞij2 larger than 0.99, where j c f i and j c ðtÞi are
the ideal final state and the real state of the system at time t,
respectively, when considering the full Hamiltonian
without rotating-wave approximation, and the effect of
the other passive phonon modes in the trap (see below).
The ion-to-phonon transfer is standard in trapped ion technology [19] and can be done with large fidelity: The carrier
Stark shift can be canceled by detuning the laser, and the
effect of the other modes is negligible for this Rabi frequency. Individual ion addressing is easy in this regime.
For step II, in order to couple the phonon to a collective
internal ion excitation j0; npn ¼ 1i ! j1; npn ¼ 0i, we assume that states j1i and j0i are connected by a metastable
quadrupole electric transition. The reason for considering
this kind of transition here is that, as opposed to step III
above, there is a phase matching condition: the collective
excitation has to fit the standing wave pattern of the final
photon in the cavity for optimal coupling, due to the ionic
inhomogeneous spacing. At first glance, a simple Raman
transition would be the easiest choice here. However,
because of the need to align the coupling laser(s) optimally
with the modes, we find a quadrupole transition made from
two pairs of lasers works best [depicted in Fig. 1(a)]. Thus
we consider a slightly noncoaxial standing laser field configuration, with two pairs of lasers pairwise opposite in the
x direction, with one pair pointing rightwards and the other
one pointing leftwards in the z direction; see Fig. 1(a).
The four lasers will give a joint field of
Aðx; z; tÞ / cosðkx xÞ sinðkzÞ cosð!0 tÞ;
(1)
i.e., for all the ions with x ¼ 0, this gives sinðkzÞ cosð!0 tÞ
with a correction in x fluctuations that is quadratic in the
Lamb-Dicke parameter, and thus can be neglected. These
lasers will have frequency !0 , resonant with the quadrupolar transition. The quadrupolar Hamiltonian contains the
gradient of the field, such that the resulting Hamiltonian
will have a cosðkzÞ dependence and will exhibit the typical
standing wave sinusoidal pattern needed, matching the
cavity mode: cosðkzi Þ ’ cosðkz0i Þ kzi sinðkz0i Þ, where
zi is the fluctuation of the position operator zi around
the equilibrium position, z0i . This way, the ions will couple
to the phonon with equal relative couplings as to the
photon, thus maximizing the fidelity. Also, due to the angle
the lasers make with the cavity axis, their frequency will
correspond to the quadrupole electric transition, which
usually is different from the cavity frequency. The resulting
complete Hamiltonian reads
pffiffiffi
X
H1 ¼ !bypn bpn þ 3!b~ypn b~pn þ ðþ
i þ i Þ
i
cosðkz0i Þ sinðkz0i Þ pffiffiffiffi ðbpn þ bypn Þ
N
X
~ ~
þ pffiffiffiffi ðbpn þ b~ypn Þ þ j1ii h1j;
N
i
(2)
where bypn is the main phonon mode, e.g., the c.m. mode,
that we will use for the protocol, and the single-ion
Lamb-Dicke parameter. In addition, the effect of the other
modes will be summed up in b~ypn , whose frequency we take
pffiffiffi
as 3! (the stretch mode frequency, nearest to the c.m.
one), by considering a larger Lamb-Dicke parameter for
this mode. Our estimates indicate that choosing ~ ¼ 0:4 ¼
4 is conservative enough to slightly overestimate the
030501-2
PRL 107, 030501 (2011)
spurious effect of the other modes. Accordingly, this is the
value we will consider. The other quantities are max ¼
0:01!, where ðtÞ has a Gaussian profile, and ¼ 0:1.
This operation is optimal if one performs an adiabatic
sweep of the detuning over the resonance with respect
to the red sideband of the j0i ! j1i transition as plotted in
Fig. 1(c). We take a maximum value of the detuning of
j !j ¼ 8 103 !, and we consider a chirp with linear dependence on time. The process fidelity is given by
F1 ¼
Z
week ending
15 JULY 2011
PHYSICAL REVIEW LETTERS
ddh c f jTr½Uj c i ih c i jUy j c f i
ð1 jj2 jj2 Þ;
where j c i i ¼ j0; npn ¼ 0i þ j0; npn ¼ 1i and j c f i ¼
j0; npn ¼ 0i þ j1; npn ¼ 0i, U is the evolution operator
associated with H1 , and the trace is taken over the spurious
mode. F1 is depicted in Fig. 2(a) as a function of the ion
number N. While the main phonon mode, the c.m. mode,
should be cooled down to the ground state for performing
the protocol, the remaining modes in principle may be
cooled just to the Doppler limit. We verified that there is
no transfer of population from states j0;npn ¼ 1i and
j1;npn ¼ 0i to other spurious phonon modes, given that they
are far off-resonant. In addition, the Stark shifts induced by
the spurious modes are negligible at the Doppler limit.
Accordingly, these modes need not be cooled down to the
ground state. For the matching to the cavity mode to work,
the quadrupolar interference pattern needs to be interferometrically stable with respect to the cavity standing wave.
Finally, step I consists of coupling the collective ion
excitation to the cavity mode ay , that subsequently will
exit the cavity [23]. This can be done using a straightforward Raman transition in a STIRAP setup, between levels
j1; nph ¼ 0i and j0; nph ¼ 1i [see Fig. 1(b)]. The 1 laser
is directed perpendicularly to the chain, such that no additional phases are introduced. The Hamiltonian and master
equation for this system read
X
H2 ¼ ½1 ðjeii h1j þ j1ii hejÞ þ g0 sinðkz0i Þ
i
X
ðjeii h0ja þ H:c:Þ þ jeii hej;
(3)
i
_ ¼ i½H2 ; þ ð2aay ay a ay aÞ
2
X
þ ðj1ii hejjeii h1j þ j0ii hejjeii h0j
i
jeii hej jeii hejÞ;
(4)
with cavity decay rate . We can describe the transition
using a STIRAP [22] pulse. The fidelity F2 of this process
will be dominated by spontaneous emission from jei
F2 ¼ 1 2
FIG. 2 (color online). (a) Process fidelity F1 for the transfer of
one phonon onto one collective excitation versus the number of
~¼
ions N. The parameters used are max ¼ 102 !, ¼ 0:1, 0:4. The line is an average to guide the eye. (b) Fidelity F2 for
the transfer of one collective excitation to one photon versus the
number of ions N. The parameters used are 1 ¼ 50, ¼ 10,
¼ 0. The lines are averages to guide the eye.
Z1
0
X
dt hei jðtÞjei i:
(5)
i
We plot F2 in Fig. 2(b) versus the number of ions N.
Note that F2 increases towards 1 with the number of ions
N. This growth is especially significant for low N. For
!=2 ¼ 1 MHz, the total transfer time of steps I–III is of
about 2.3 ms for 20 ions. For the joint process fidelity of
steps I–III, we obtain an optimal value of 0.98 for N ¼ 18
and g0 ¼ 8.
Having explained the basic setup, we can now discuss
some potential extensions and applications. One such extension, two or more excitations, can be done in a straightforward manner. In this case, we consider a superposition
state with n photons 0 j0i þ 1 j1i þ þ n jni which
is mapped to the internal state 0 j01 02 . . . 0N iþ
1 j01 . . . 1i1 . . . 0N i þ þ n j01 . . . 1i1 . . . 1in . . . 0N i,
where i1 through in denote n definite (potentially neighboring) ions. The mapping process can then be achieved much
the same way as described above, but the ‘‘single’’ ion-tophonon step needs n ions to be addressed individually by
one or more lasers. In the phonon stage, jni corresponds to
npn ¼ n, in the collective-ion-excitation stage to the Dicke
state with n excitations. The collective enhancement factor
increases in this case to ðNnÞ1=2 . For the coherent transfer
from two excitations to two photons, for example, we
obtain an F2 (which is a lower bound to the probability
of no spontaneous emission in the two-excitation case)
030501-3
PRL 107, 030501 (2011)
PHYSICAL REVIEW LETTERS
from Eq. (5) of 0.97 for 12 ions, with the parameters 1 ¼
50, ¼ 0, ¼ 10, g0 ¼ 8.
An important practical consideration for the implementation of any such operations concerns mapping of incident
photon states into the ion crystal. In the ideal limit, perfect
mapping can be achieved, via time reversal of the spinphoton mapping procedure discussed above. In practice,
the fidelity of this process will be limited by finite optical
depth, i.e., by the number of ions in the crystal. The
procedure of coupling photons into ensembles without
and with cavities has been investigated in detail previously
[17,18,24]. In general, for optimally chosen coupling strategies, the storage efficiencies will be similar to the retrieval
efficiencies [14,15].
There are multiple options for this scheme to be used as
a photonic gate, i.e., a nonlinearity that acts on two or more
photons. In the general case we will consider a superposition j c ph i ¼ 0 j0i þ 1 j1i þ 2 j2i of 0, 1, and 2 photons
for the incoming state, and the aim will be to introduce a
minus sign in the two-photon state, to get Uð2Þ j c ph i ¼
0 j0i þ 1 j1i 2 j2i. A way to achieve that is based on
transfer of the state j c ph i to a superposition of 0, 1, and 2
collective excitations, j c col i ¼ 0 j0i þ 1 j1i þ 2 j2i,
where j2i is thePcollective Dicke-like state with two excitations, j2i / ij gi gj j01 02 . . . 1i . . . 1j . . . 0N i. Transferring subsequently to the superposition of zero-, one-, and
two-phonon states, j c pn i ¼ 0 jnpn ¼ 0i þ 1 jnpn ¼ 1i þ
2 jnpn ¼ 2i, and from it to the same superposition of
single-ion states, one would get 0 j01 02 . . . 0N i þ
1 j01 . . . 1i1 . . . 0N i þ 2 j01 . . . 1i1 1i1 þ1 . . . 0N i. We propose then to perform a standard 2-qubit phase gate upon
the internal states of ions i1 and i1 þ 1, introducing a minus
sign upon the 2 j01 . . . 1i1 1i1 þ1 . . . 0N i state. For example,
the quantum-bus [25] or Sørensen-Mølmer [26] gates
could be implemented in a straightforward manner.
Subsequently, one would undo all the previous steps, and
retrieve the photonic state with the minus sign incorporated
upon the 2 jnph ¼ 2i component of the state, thus completing the two-photon gate. A further possibility could be
to consider the change in absorption of the cavity produced
by the presence or absence of an excitation in the ions
inside the cavity [27]. Thus, the phase of an incoming
photon would be conditionally modified in the case where
a former photon was previously absorbed.
From the above descriptions it is obvious that there is a
large number of applications possible with this scheme.
This includes quantum gates and quantum nondemolition
measurements of single-photon qubits [2], efficient entanglement purification in ion-based quantum repeater
schemes [28], efficiency enhancement in probabilistic ion
entanglement schemes [29], as well as deterministic quantum gates between distant trapped ions [23].
Finally, we point out that the here envisioned level
scheme may be attained in several presently possible ion
trap setups such as 40 Caþ or 88 Srþ trapped ions.
week ending
15 JULY 2011
We thank M. C. Bañuls, A. V. Gorshkov, O. RomeroIsart, T. Schätz, and H. Schwager for fruitful discussions.
We acknowledge funding from EU projects AQUTE and
COMPAS, from DFG Forschergruppe 635, from the NSF
CUA, DARPA QUEST, AFOSR MURI, and the Packard
Foundation. S. F. Y. acknowledges funding from NSF
through Grant No. PHY-0970055.
*lucas.lamata@mpq.mpg.de
[1] M. Nielsen and I. Chuang, Quantum Computation and
Quantum Information (Cambridge University Press,
Cambridge, England, 2000).
[2] E. Knill, R. Laflamme, and G. J. Milburn, Nature
(London) 409, 46 (2001).
[3] M. D. Lukin, S. F. Yelin, and M. Fleischhauer, Phys. Rev.
Lett. 84, 4232 (2000).
[4] M. Fleischhauer, S. F. Yelin, and M. D. Lukin, Opt.
Commun. 179, 395 (2000).
[5] M. Fleischhauer and M. D. Lukin, Phys. Rev. Lett. 84,
5094 (2000).
[6] A. E. Kozhekin, K. Mølmer, and E. Polzik, Phys. Rev. A
62, 033809 (2000).
[7] S. A. Moiseev and S. Kröll, Phys. Rev. Lett. 87, 173601
(2001).
[8] A. Kuzmich et al., Nature (London) 423, 731 (2003).
[9] D. N. Matsukevich and A. Kuzmich, Science 306, 663
(2004).
[10] A. T. Black, J. K. Thompson, and V. Vuletic, Phys. Rev.
Lett. 95, 133601 (2005).
[11] B. Kraus et al., Phys. Rev. A 73, 020302(R) (2006).
[12] J. Simon et al., Phys. Rev. Lett. 98, 183601 (2007).
[13] A. V. Gorshkov et al., Phys. Rev. Lett. 98, 123601
(2007).
[14] A. V. Gorshkov et al., Phys. Rev. A 76, 033804
(2007).
[15] A. V. Gorshkov et al., Phys. Rev. A 76, 033805
(2007).
[16] P. F. Herskind et al., Nature Phys. 5, 494 (2009).
[17] A. I. Lvovsky, B. C. Sanders, and W. Tittel, Nat. Photon. 3,
706 (2009).
[18] C. Simon et al., Eur. Phys. J. D 58, 1 (2010).
[19] D. Leibfried et al., Rev. Mod. Phys. 75, 281 (2003).
[20] A. B. Mundt et al., Phys. Rev. Lett. 89, 103001 (2002).
[21] M. Keller et al., Appl. Phys. B 76, 125 (2003).
[22] K. Bergmann, H. Theuer, and B. W. Shore, Rev. Mod.
Phys. 70, 1003 (1998).
[23] J. I. Cirac et al., Phys. Rev. Lett. 78, 3221 (1997).
[24] C. W. Gardiner and P. Zoller, Quantum Noise (SpringerVerlag, Berlin, 2004).
[25] J. I. Cirac and P. Zoller, Phys. Rev. Lett. 74, 4091
(1995).
[26] A. Sørensen and K. Mølmer, Phys. Rev. Lett. 82, 1971
(1999).
[27] L.-M. Duan and H. J. Kimble, Phys. Rev. Lett. 92, 127902
(2004).
[28] H.-J. Briegel et al., Phys. Rev. Lett. 81, 5932
(1998).
[29] C. Cabrillo et al., Phys. Rev. A 59, 1025 (1999).
030501-4